pipecat-aipipecat-aiOfficialยท1 Agent Skills Included

pipecat-mcp-server

Voice conversations and screen capture for coding agents

Adds real-time voice conversations and screen capture to coding agents through a local MCP server. Runs speech-to-text and text-to-speech locally with Whisper and Kokoro, so no API keys are required. Connects through browser, WebRTC rooms, or phone calls, and lets the agent analyze your screen to debug errors or review designs.
npx skills add pipecat-ai/pipecat-mcp-server --all -g -y

All Skills in This Repository (1)

Pure Emerald Level Indicators

Frequently Asked Questions

FAQPage Schema
How to install Pipecat MCP Server?โ–ผ

Run `npx skills add pipecat-ai/pipecat-mcp-server --all -g -y` in your terminal to install everything globally.

How to talk to my coding agent by voice?โ–ผ

Start the server with `pipecat-mcp-server`, open the local playground in your browser, and speak naturally. The agent hears you, responds by voice, and can run commands on your machine.

Does Pipecat MCP Server need API keys?โ–ผ

No. By default it uses local Whisper for speech-to-text and Kokoro for text-to-speech, so it works fully offline without any API keys.

Does it work with Claude Code, Cursor, and Codex?โ–ผ

Yes. It follows the MCP standard, so any compatible client can register the server at http://localhost:9090/mcp and start voice sessions.

Can the agent see my screen?โ–ผ

Yes. On macOS and Linux you can stream your screen or a specific window, then ask the agent to debug errors or review your UI based on what it sees.

Related Repositories in Software Engineering

View All in Software Engineeringโ†’